The subject matter disclosed herein relates to a time of flight, TOF, camera unit for an optical surveillance system and to an optical surveillance system comprising such a TOF camera.
During the last few years, the first all-solid state three-dimensional cameras (3-D cameras) based on the time of flight (TOF) principle became available on the market. 3-D cameras or sensors based on the time of flight principle acquire distance information from objects in a scene being imaged. Distance information is produced independently at each pixel of the camera sensor. As for instance described in U.S. Pat. No. 6,323,942, a TOF system emits optical energy and determines how long it takes until at least some of that energy reflected by a target object arrives back at the system to be detected. Emitted optical energy traversing to more distant surface regions of a target object before being reflected back towards the system will define a longer TOF than if the target object were closer to the system. If a round trip TOF is denoted t, then the distance d between the target object and the TOF system can be calculated as d=t·c/2, with c being the velocity of light.
Such known systems can acquire both luminosity data (signal amplitude) and TOF distance, and can produce three-dimensional images of a target object in real time.
Rather than directly measuring a time of flight, which requires very sensitive imaging equipment, a more sophisticated principle is based on a phase measuring TOF principle. Here, instead of directly measuring a light pulse's total trip, the phase difference between sent and received signals is determined. When modulating the transmitted light with a modulation frequency, fm, the distance between the point of reflection and the camera can be calculated as
                    d        =                              c                          2              ⁢                              f                m                                              ·                      φ                          2              ⁢              π                                                          (        1        )            
The detection of the reflected light signals over multiple locations in a system pixel array results in measurement signals, that are referred to as depth images. The depth images represent a three-dimensional image of the target object surface.
FIG. 4 shows such a known system 200 for surveillance of a predefined surveillance area. As shown in FIG. 4, a TOF camera 202 is mounted in a distance from a background 204 that allows to survey a sufficiently large detection field. In the figure, only the radiation beams emitted by the TOF camera 202 are schematically shown and denoted with reference numeral 206, but not the returning detected beams. Depending on the aperture of the employed TOF camera 202, the detection field is limited as indicated by detection field limits 208.
A problem arises when an object to be detected 210 enters the detection field. As schematically shown in FIG. 5, those radiation beams 206, which reach the upper surface of the object 210, are reflected and reach the TOF camera 202, providing a three-dimensional image of the first surface 212 of the object 210. However, due to shading effects with the known surveillance system 200 the problem occurs that invisible areas 214 of the object 210 remain. For an unambiguous automatic identification of the object 210, it is often necessary to also receive information about these normally invisible areas 214.
In order to solve this problem, a first possible solution is to provide two cameras 202 and 203. A first camera 202 is oriented in a way that it can monitor area 216 and a second camera 203 is arranged in a way that it can monitor the area 218, which is invisible for camera 202. By combining the images of the two cameras according to the system shown in FIG. 6, the contour of an object can be detected unambiguously. However, the system is costly, requiring two complete TOF cameras and the synchronization of the generated images is difficult.
For an object contour monitoring for larger areas, for instance a door or a through way, it is known from U.S. Pat. No. 5,903,355 to use mirrors in combination with a laser scanner, in order to allow for an indirect detection of shapes which are not visible to the scanner directly, by means of the reflection of the scanning ray in the mirrors. To this end, a larger deviation angle of the laser beam than need for the viewing area must be scanned, so that the mirror area is included in the scanning movement as well. However, as, this is known for laser scanners, the viewing field of this arrangement is only in one plane and no three-dimensional image can be generated from the retrieved data. Furthermore, the signal processing routine proposed in this document is only working in combination with a moving laser scanning beam. Furthermore, the arrangement according to U.S. Pat. No. 5,903,355 is not suitable for short-distance applications.
Consequently, there exists a need for an improved optical surveillance system which on the one hand is robust, economic and user friendly and which on the other hand yields comprehensive three-dimensional contour information also for short distanced objects.